The Wrong Question About the Tech Market

There's a question that keeps coming up in every hallway conversation, every coffee with colleagues, every message that lands in my inbox these days: "How many more years until the tech market stabilizes?"

I get it completely. After two years of layoff headlines, people want a date, a point on the horizon to plant their feet on. But the more I hear it, the more convinced I am that it's the wrong question. I want to walk through why with you, step by step, because at the end of this path I don't want to leave you with a worry — I want to leave you with an open door.

Let's start with the trap. It's hiding in the word "stabilize," which assumes the market is a single wave that rises and falls, and that all we have to do is wait for the tide to go out. But that's not what the 2026 data shows. What's happening isn't the market destabilizing — it's splitting in two, and each half is running in the opposite direction.

On one side, generic software development remains close to 49% below its pre-pandemic level. On the other, AI and machine learning roles sit around 59% above that same baseline (Indeed Hiring Lab). Same sector, same year, and yet these are two realities that barely touch each other.

That's exactly why the layoff headlines are so deceiving. They're not the market collapsing — in large part, they're reassignment. The same companies cutting roles are hiring aggressively in other functions. A layoff announcement from a healthy company rarely means the business is contracting; it means one type of role is being replaced by another. And if that's true, the question stops being "when does it go back to how it was" and becomes "which direction is the ground actually moving."

This is where the path gets interesting, because there's one number that explains almost everything. 95% of enterprise AI pilots never make it to production (MIT NANDA study). And the revealing part is that almost none of those failures come down to the model. They come down to deployment — nobody managed to embed that technology inside the real, living, messy system of an actual company. That 95% isn't a depressing headline. It's a massive gap waiting for someone to fill it.

Out of that gap came the year's most talked-about role: the Forward Deployed Engineer, the FDE. The engineer who leaves the office and steps fully into the client's organization to build, adapt, and make AI actually work in the real world, with all its constraints and ambiguity. a16z called it, without much exaggeration, "the hottest job in tech," and its openings have grown by hundreds of a percent in barely a year.

And now, here's the part I want you to walk away with, because this is where this stops being analysis and becomes an invitation.

The FDE role doesn't reward whoever memorized last quarter's trendy framework. It rewards a combination that's far more human — and far more learnable — than it looks. It rewards breadth: writing backend one day, an integration the next, a data pipeline the day after that. It rewards knowing how to talk to the business side and translate technical work for people who aren't technical. And it rewards a new territory that anyone can start exploring today: understanding how AI agents work, RAG, vector databases, and above all the discipline that truly separates a demo from a real deployment — evals and guardrails, knowing how to prove an AI system won't spiral out of control in production.

None of that requires starting from zero. It builds on top of what many of us already know. And that's the good news for those of us who've spent years in complex enterprise environments: those years we sometimes feel are a disadvantage against recent graduates are, in this market, exactly the asset that's in short supply. Demand for these profiles is growing exponentially while the supply barely grows linearly. For whoever decides to prepare today, that gap is the best window we've had in a long time.

So the honest question is no longer "when do I wait for the market to get better?"

It's one you can actually answer yourself, starting this very week: "what am I going to learn, what am I going to explore, to stand in the half of the market that's growing?"

I'd love to hear from you. What are you seeing from your own trench: reassignment or contraction? And if you had to start exploring just one of these skills this month, which one would you pick?